Abstract

   Some effects of aging in animals are tissue-specific. In D.
   melanogaster neuronal overexpression of Gclc increases lifespan and
   improves certain physiological parameters associated with health
   benefits such as locomotor activity, circadian rhythmicity, and stress
   resistance. Our previous transcriptomic analyses of Drosophila heads,
   primarily composed of neuronal tissue, revealed significant changes in
   expression levels of genes involved in aging-related signaling pathways
   (Jak-STAT, MAPK, FOXO, Notch, mTOR, TGF-beta), translation, protein
   processing in endoplasmic reticulum, proteasomal degradation,
   glycolysis, oxidative phosphorylation, apoptosis, regulation of
   circadian rhythms, differentiation of neurons, synaptic plasticity, and
   transmission. Considering that various tissues age differently and
   age-related gene expression changes are tissue-specific, we
   investigated the effects of neuronal Gclc overexpression on gene
   expression levels in the imago thorax, which is primarily composed of
   muscles. A total of 58 genes were found to be differentially expressed
   between thoraces of control and Gclc overexpressing flies. The Gclc
   level demonstrated associations with expression of genes involved in
   the circadian rhythmicity, the genes in categories related to the
   muscle system process and the downregulation of genes involved in
   proteolysis. Most of the functional categories altered by Gclc
   overexpression related to metabolism including Drug metabolism,
   Metabolism of xenobiotics by cytochrome P450, Glutathione metabolism,
   Starch and sucrose metabolism, Citrate cycle (TCA cycle), One carbon
   pool by folate. Thus, the transcriptomic changes caused by
   neuron-specific Gclc overexpression in the thorax were less pronounced
   than in the head and affected pathways also differed from previous
   results. Although these pathways don't belong to the canonical
   longevity pathways, we suggest that they could participate in the delay
   of aging of Gclc overexpressing flies.

   Keywords: Gclc, glutathione, lifespan, aging, gene expression, thorax,
   Drosophila melanogaster

Introduction

   The aging process is determined by the impairment functioning of
   tissues and systems of an organism. It was shown that aging is
   accompanied by significant changes in genome-wide gene expression
   profiles identified for different tissues of vertebrates and
   invertebrates (Zhan et al., [39]2007; Kim et al., [40]2014; White et
   al., [41]2015; Cannon et al., [42]2017; Chen et al., [43]2018; Zhang et
   al., [44]2018). Furthermore, the age-related changes in gene expression
   are tissue-specific, suggesting that various tissues age differently
   (Girardot et al., [45]2006; Zhan et al., [46]2007; Yang et al.,
   [47]2015). Recently, a number of studies showed that muscle tissue
   undergoes sufficient alterations in gene expression patterns during
   aging in mammals and flies (Kim et al., [48]2014; Chen et al.,
   [49]2018; Zhang et al., [50]2018; Zhou et al., [51]2018), and revealed
   the involvement of several tens of messenger RNAs as well as long
   non-coding, micro, and circular RNAs in this process (Kim et al.,
   [52]2014; Chen et al., [53]2018). In particular, aged rats demonstrated
   decreased expression of components of AMPK, IGF-1, and CASK pathways
   that can be a key cause of the loss of muscle mass and function (Zhou
   et al., [54]2018). The progressive loss of skeletal muscle function and
   mass, known as sarcopenia, is one of the useful conditions accompanied
   the aging process in human and underlied the critical mobility reducing
   and life quality decline (Herndon et al., [55]2002; Demontis et al.,
   [56]2013; Larsson et al., [57]2019). It must be noted, that the
   aging-related muscle degeneration are also determined by the
   progressive loss of motoneurons (Larsson et al., [58]2019). Therefore,
   interventions that contribute to the slowing down of aging of nervous
   tissue (including overexpression of pro-longevity genes) can also have
   a positive effect on the state of muscle tissue and prevent dangerous
   structural and cellular changes.

   According to the recent experimental data, Drosophila is a convenient
   and relevant model for the investigation of muscle degeneration
   disorders (Demontis et al., [59]2013; Rai et al., [60]2014; Kreipke et
   al., [61]2017). Despite differences in the composition of mammalian and
   fly muscle systems, the Drosophila model can be used for the
   identification of key molecular mechanisms of aging-related changes in
   muscle tissue (Demontis et al., [62]2013; Rai et al., [63]2014).
   Furthermore, the lack of stem cell-based regeneration allows to
   determine the role of endogenous processes of muscle degeneration
   (Kreipke et al., [64]2017). The study of molecular basis of skeletal
   muscle laminopathies were carried by using of Drosophila melanogaster
   LamC mutants which demonstrated a numbered of muscular pathologic
   changes and cellular abnormalities. These negative changes were
   prevented by regulation of AMP-activated protein kinase
   (AMPK)/TOR/autophagy signaling pathways that highly conserved between
   flies and humans (Chandran et al., [65]2018). Thus, the study of
   mechanistic basis of muscular disorders using Drosophila allows to
   reveal potential targets for therapeutic interventions.

   The Gclc is a target gene of pro-longevity SKN-1/Nrf2 transcription
   factor, activated in response to oxidative stress, and implicated in
   the lifespan extension of flies, worms, and mice (Orr et al., [66]2005;
   Luchak et al., [67]2007; Sykiotis et al., [68]2011; Mockett and Nobles,
   [69]2013; Blackwell et al., [70]2015). It codes for the catalytic
   subunit of glutamate-cysteine ligase, that participates in glutathione
   formation (Orr et al., [71]2005). Age-dependent downregulation of
   glutamate-cysteine ligase is associated with a decline in glutathione
   content, which is involved in multiple cellular functions including
   antioxidant defense and detoxification of reactive intermediates (Liu
   et al., [72]2004; Orr et al., [73]2005). Previously, it was
   demonstrated that the neuron- and muscle-specific overexpression of
   Gclc is associated with increased lifespan in flies, but not the
   ubiquitous overexpression of Gclc (Orr et al., [74]2005; Luchak et al.,
   [75]2007; Mockett and Nobles, [76]2013). In addition, Gclc
   overexpression in the nervous system postpones age-associated decline
   in circadian rhythmicity and a variety of physiological parameters like
   locomotor activity and resistance to the oxidative, osmotic, and
   proteotoxic stresses (Moskalev et al., [77]2016). Gclc overexpression
   also affects transcription patterns of genes that govern diverse
   cellular processes such as Jak-STAT, MAPK, FOXO, Notch, mTOR, TGF-beta
   signaling pathways, translation, protein processing in endoplasmic
   reticulum, proteasomal degradation, glycolysis, oxidative
   phosphorylation, apoptosis, regulation of circadian rhythms,
   differentiation of neurons, synaptic plasticity, and transmission in
   Drosophila heads (Moskalev et al., [78]2016).

   We supposed, that neuron-specific Gclc overexpression may ameliorate
   the decline of muscular function and delay the aging of muscle tissue
   (Moskalev et al., [79]2016). For the first time, we studied the effect
   of neuronal overactivation of the pro-longevity gene on age-related
   transcriptome changes in the muscular system. We used total RNA
   sequencing and revealed that Gclc overexpression in the head also
   affected the key signaling pathways implicated in the aging process in
   the thorax, which is composed primarily of muscle tissue. We identified
   58 differentially expressed genes. Among the functional categories
   [Gene Ontology (GO) terms or Kyoto Encyclopedia of Genes and Genomes
   (KEGG) pathways] the genes consistently overrepresented were associated
   with Drug metabolism, Metabolism of xenobiotics by cytochrome P450,
   Glutathione metabolism, Starch and sucrose metabolism, Neuroactive
   ligand-receptor interaction, One carbon pool by folate, vesicle, and
   Cdc73/Paf1 complex. While apparent, the transcriptomic changes in the
   thorax were less pronounced than in the head.

Materials and methods

Drosophila melanogaster Lines

   Flies were separated by sex and maintained in different vials on a
   standard sugar-yeast medium at constant temperature (25°C) and humidity
   (60%) in a 12:12 h light-dark cycle in Binder KBF720-ICH (Binder,
   Germany) climate chamber. The following strains were used: UAS-Gclc
   (provided by Dr. William C. Orr, Southern Methodist University) and
   Appl-GAL4 (#32040, Bloomington Drosophila Stock Center). They were
   backcrossed with w^1118 (#3605, Bloomington Drosophila Stock Center,
   USA) 6–8 times to equilibrate the genetic background. To activate the
   Gclc constitutive overexpression in the nervous system, the UAS-Gclc
   was crossed with Appl-GAL4 driver. Newly enclosed progenies were
   maintained for one, four, and 6 weeks before dissection. The same aged
   flies from the parental UAS-Gclc line were used as a contro control.

Total RNA Extraction and Qualification

   The experimental design was the same as in our previous work (Moskalev
   et al., [80]2016). Thoraxes in this study and heads in the previous
   study were collected from the same flies. In brief, we used males and
   females' thorax of UAS-Gclc flies as control and thoraxes of flies with
   Gclc overexpression. A comparison of flies with different chronological
   ages, namely, 1-, 4-, and 6- weeks was conducted. The ages were
   designated as “young,” “matured,” and “old,” respectively. The flies
   were collected, frozen and kept at −80°C to prevent RNA degradation.
   For each experimental variant, 40 flies were selected. Three biological
   replicates of the experiment were performed. The males and females were
   analyzed separately to reveal any sex-dependent differences in gene
   expression caused by neuronal Gclc expression. The RNA used for
   next-generation sequencing was extracted from 30 flies (10 flies per
   each replicate) using QIAzol Lysis Reagent (Qiagen, Netherlands) and
   precipitated with isopropanol. Qubit®2.0 Fluorometer (Invitrogen, USA)
   and NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies Inc.,
   USA) were used to estimate the RNA concentration and purity. The
   A260/A280 ratio in the samples was around 1.8–2.0. RNA was additionally
   purified using DNase I (Promega, USA). The RNA integrity was estimated
   with Bioanalyzer Agilent 2100 (Agilent Technologies, USA).

Library Preparation and Sequencing Procedure

   mRNA was isolated from the total RNA with NEBNext Oligo d(T)25 beads.
   The fragmentation was performed at 94°C with addition of First Strand
   Synthesis Reaction Buffer and Random Primer mix (2×). The
   double-stranded (ds) cDNA was synthetized by using ProtoScript II
   Reverse Transcriptase and Second Strand Synthesis Enzyme Mix. The blunt
   ends on the ds cDNA were created with the end-repair reaction. The
   adapters were ligated using the specific RNA Adapter Indexes supplied
   in the kit. The PCR reaction was performed to selectively enrich
   adaptor ligated DNA fragments and create a large amount of DNA, which
   was then purified using Ampure XP beads. The quantity and quality of
   libraries were assessed on Rotor-Gene 6000 PCR System (Qiagen, USA)
   using the qPCR method and on Agilent 2100 Bioanalyzer using a High
   Sensitivity DNA chip. The final length of the libraries averaged 260
   bp. The libraries were normalized to 4 nM, pooled together in equal
   volumes, and sequenced with 2 × 50 bp paired-end reads on the NextSeq
   500 System (Illumina, USA). The obtained data were stored in FASTQ
   format. The minimum amount of reads for each sample was 20 million.

NGS Data Processing

   We used PPLine pipeline to perform quality control, read mapping, and
   counting (Krasnov et al., [81]2015). In the details, reads were
   analyzed using FastQC and trimmed with trimmomatic (Bolger et al.,
   [82]2014). To evaluate the efficacy of the isolation of polyA fraction
   and the absence of bacterial contamination, 100.000 randomly selected
   reads from each sample were mapped to D.Melanogaster rRNA genes and
   bacterial genomes (all strains that had been submitted to NCBI Genome
   through 2015), accordingly. Typically, rRNA ratio was 0.3–1.7% (except
   for 2 samples with 4%). The trimmed reads were mapped to the
   D.Melanogaster genome (assembly BDGP6, Ensembl release 90) using
   splice-aware STAR aligner (Dobin et al., [83]2013). About 90–95% reads
   were uniquely mapped. Since the samples had demonstrated different RNA
   integrity numbers (RIN), in order to compare gene expression levels, we
   take into account a 3′-tail bias as described in Sigurgeirsson et al.
   ([84]2014) with some modifications. Without a 3′-bias adjustment, one
   should expect false-positive overexpression of short transcripts (for
   example, genes of mitochondrial proteins or ribosomal proteins) and
   under-expression of genes with long transcripts and encoded proteins
   (for example, multi-domain transmembrane proteins interacting with
   extracellular matrix).

   First, we analyzed read coverage across transcript length using a
   modified version of geneBody_coverage.py script, part of RSeQC package
   (Wang et al., [85]2012). In the samples with lower RIN numbers and more
   degraded RNA, we observed rapid falling down of read coverage level
   after 750–1,000 bp from the 3′-tail. The coverage in the region of
   0–500 bp from the 3′-tail was almost preserved for all samples.

   Then, we quantified transcripts using RSEM (Li and Dewey, [86]2011) and
   identified the most abundantly expressed alternative transcript for
   each gene. The other transcripts were discarded. Then, the transcripts
   were truncated to the length of 500 bp from the 3′-end and a new GTF
   file was generated. Finally, using this gene model, we quantified reads
   using Counts from Subread package (Li and Dewey, [87]2011). This
   procedure allowed us to significantly reduce the dependence of the
   observed expression level fold changes (FC) on the transcript lengths,
   between the samples with higher and lower RINs.

   In order to more effectively eliminate the factor of the 3′-bias, we
   proceeded the following way. First, we excluded genes with low
   expression level by introducing a threshold: CPM > 2 for at least 3
   samples. About 9,000 genes passed the threshold. Next, we split all the
   genes into 10 bins, depending on the average gene transcript length,
   and within each bin we normalized read counts using the TMM (trimmed
   mean of M-values) method from edgeR (Robinson et al., [88]2010).
   Finally, the bins were merged.

   A similar procedure was performed to eliminate the dependency of the
   expression level FC on the “absolute” gene expression level (in terms
   of read counts per million, CPM). This bias may also occur when
   comparing samples with different RINs and initial RNA concentration.

   Finally, the adjusted read pseudo-counts were analyzed by edgeR using
   exact test and the quasi-likelihood ratio F-test (Robinson et al.,
   [89]2010). Gene Ontology, KEGG, and Reactome enrichment was performed
   using topGO and clusterProfiler Bioconductor packages (Yu et al.,
   [90]2012). For this purpose, we used Fisher exact test and lists of top
   50, 100, 200, 500, and 1,000 up- or down-regulated genes that passed p
   < 0.05 threshold. Fisher test was applied for each of these top lists,
   and then the results were merged. As a background, we took only genes
   that passed the pre-filtering by the expression level. KEGG pathways
   visualization was performed using modification of pathview package (Luo
   and Brouwer, [91]2013) as described earlier (Moskalev et al.,
   [92]2016).

Reverse Transcription and qPCR Analysis

   The total RNA isolated for the RNA-Seq library preparation was used for
   qPCR analysis. First-strand cDNA was synthesized using 0.5 μg of total
   RNA, random primers and Mint reverse transcriptase (Evrogen, Russian
   Federation). Real-time quantitative PCR reactions were performed in
   triplicate in the presence of the EvaGreenTM dye (Biotium Inc., USA)
   and gene specific primers ([93]Table 1) on the ABI 7500 Fast Real-Time
   PCR System (Applied Biosystems, USA). The cycling parameters were 95°C
   for 10 min, followed by 40 cycles of 95°C for 15 s, 60°C for 60 s, and
   72°C for 30 s. act5c and tub84B were used as reference genes. The
   relative expression ratios were calculated using the 2–ΔΔCt method as
   previously described (Dmitriev et al., [94]2016). The Wilcoxon and
   Mann-Whitney tests were applied for assessment of differences in mRNA
   expression. P ≤ 0.05 were considered statistically significant.

Table 1.

   Nucleotide sequences of primers used for qPCR.
   Gene Forward primer (5′-3′) Reverse primer (5′-3′) Product length (in
   bp)
   Mlp60A cgggtagaaccttcgtgaga gtgcgaggcatgcaagattt 73
   Act88F ctcggctcggacagtgatag gctataccgctgccgatgaa 144
   Arc1 catcatcgagcacaacaacc ctactcctcgtgctgctcct 169
   act5c gagcgcggttactctttcac acttctccaacgaggagctg 133
   tub84B aacctgaaccgtctgattgg ggtcaccagagggaagtgaa 134
   [95]Open in a new tab

Results

   In this study, we analyzed the effects of Gclc overexpression in the
   neuronal tissues of D. melanogaster on gene expression profiles in the
   thorax. We derived RNA-Seq expression profiles for 9,000 genes (after
   eliminating the genes with low expression). Drosophila aging was
   followed by the differential expression of 2626 genes (p < 0.05). 1488
   of 2626 passed FDR < 0.05 threshold. The expression of 667 genes (p <
   0.05) was associated with the activation of Gclc in all groups:
   young/mature/old or males/females (48 of 667 have FDR < 0.05). 58 of
   667 genes demonstrated 2-fold or higher expression level changes
   ([96]Supplementary Table 1). Using the STRING (Search Tool for
   Genesis/Proteins database) we received some information on the
   interactions of proteins encoded by these genes and showed that the
   most enriched pathway is the defense response to Gram-positive
   bacterium (FDR = 0.0019; [97]Figure 1). As can be seen, the genes
   included in the proteolysis GO category (highlighted in green) also
   form a cluster. It is known that there is an increase in proteolysis of
   myofibril proteins during muscular dystrophy (Costelli et al.,
   [98]2005; Combaret et al., [99]2009). According to the results of the
   DE analysis, overexpression of Gclc in thorax repressed genes involved
   in the proteolysis process.

Figure 1.

   [100]Figure 1
   [101]Open in a new tab

   Network demonstrating associations (protein interactions,
   co-expression, etc.) between 58 genes, expression of which is altered
   in transgenic flies (overexpressing Gclc; STRING database). Clusters
   formed by genes are marked different colors.

   We also verified the expression levels of some muscle-related genes
   using the qPCR method ([102]Supplementary Table 2 and
   [103]Supplementary Figure 1), the results showed a similar change in
   expression of Mlp60A, Act88F, and Arc1 genes compared with the DE
   analysis.

   The survival analysis of the examined flies demonstrated that the
   increase of median and maximum lifespan of transgenic Drosophila is
   more pronounced in females than in males (Moskalev et al., [104]2016).
   In this study, we compared the resulting DE genes list (old vs. young
   comparisons for males and female's groups) as depicted in the Venn
   diagram of [105]Figure 2, which resulted in 19 common DE genes shared
   by all 4 lists, 25 DE genes shared by Tg males and Tg females. There
   are only 6 genes common in NTg males and NTg females lists of DE genes.
   These genes are presented in [106]Supplementary Table 3.

Figure 2.

   Figure 2
   [107]Open in a new tab

   Venn diagram illustrating the results of differential gene expression
   analysis (old vs. young flies) in various groups of samples.
   Tg-transgenic, Gclc-overexpressing; NTg-non-transgenic. In Venn
   diagram, we included differentially expressed genes that passed
   thresholds: fold change > 4 and p < 0.05.

   KEGG pathway enrichment analysis suggested that the activation of
   transgenic Gclc is followed with the differential expression of genes
   involved in a variety of cellular processes including Drug metabolism,
   Metabolism of xenobiotics by cytochrome P450, Glutathione metabolism,
   Starch and sucrose metabolism, Neuroactive ligand-receptor interaction,
   One carbon pool by folate, vesicle, and Cdc73/Paf1 complex (Fisher test
   p < 0.05).

   Considering age-associated expression changes, we found predominant
   up-regulation of genes participating “One carbon pool by folate”
   pathway and a slight down-regulation of genes involved in Porphyrin
   metabolism and Neuroactive ligand-receptor interaction in Appl-GAL4 >
   UAS-Gclc flies ([108]Supplementary Figure 2). For UAS-Gclc flies, we
   found aging-associated up-regulation of Genetic Information Processing,
   including the RNA polymerase, Ribosome, Protein export, and Protein
   processing in endoplasmic reticulum, whereas in the experimental group
   these changes were not observed ([109]Figure 3 and [110]Supplementary
   Figure 2).

Figure 3.

   [111]Figure 3
   [112]Open in a new tab

   Differential expression profiles for genes participating in cellular
   pathways for different 641 genotypes (UAS-Gclc flies designated as Tg,
   Appl-GAL4 > UAS-Gclc flies—as NTg) and ages (old, mature, young).
   Logarithmic Fold Change (LogFC) (vertical axis) is ranged from −2 to
   +2, i.e., from four-fold decrease (blue) to 4-fold increase (red). Cell
   borders demonstrate the statistical significance of gene set enrichment
   analysis (Fisher test p-value): blue—enriched with down—regulated
   genes; and red—enriched with overexpressed ones min p-value across
   tests (up)—minimal enrichment p-value (Fisher exact test) across tests,
   only for up-regulated genes; min p-value across tests (down)—minimal
   enrichment p-value (Fisher exact test) across tests, only for
   down-regulated genes. DE genes that passed threshold LogCPM > 3 were
   presented.

   Using the Reactome pathway database, we also identified several
   pathways enriched with genes, the expression of which is altered by
   Gclc overexpression. Among them: response to CaM pathway, DAG and IP3
   signaling, Amino acid synthesis and interconversion (transamination),
   Peroxisomal lipid metabolism, Pyrimidine, and Purine catabolism
   ([113]Supplementary Table 4).

Discussion

   Overall, our data demonstrate that the neuronal Gclc overexpression
   resulted in less pronounced effects on the gene expression, the key
   signaling pathways involved in the aging process in the thorax than in
   the head. For example, the differential expression (DE) analysis
   revealed that only 58 genes (|LogFC|>1, p < 0.05) were differentially
   expressed in the thorax, while in the head, the neuronal Gclc
   overexpression altered the expression level of 188 genes. In heads, the
   Gclc overexpression influenced the following pathways: oxidative
   phosphorylation, ribosome biogenesis, Wnt, mTOR/PI3K, FOXO pathways,
   and enhanced cAMP signaling (Moskalev et al., [114]2016). In the
   thorax, no statistically significant effects of the neuronal Gclc
   overexpression were observed.

   Only two genes: cytochrome P450 4p2 (Cyp4p2) and white (w), were
   differentially expressed in both the head and in the thorax. The Cyp4p2
   gene encodes a member of the cytochrome P450 superfamily of enzymes.
   The w gene, famous for its role as the importer of eye pigments
   precursor, also plays a role in memory and cGMP transport (Evans et
   al., [115]2008; Sitaraman et al., [116]2008). The null mutation of this
   gene in D. melanogaster was associated with decreased lifespan,
   stress-resistance, and locomotor activity (Ferreiro et al., [117]2017).
   Using the GML model “~Gender + Genotype,” we indicated that the effect
   of Gclc expression on the transcriptional profile of the thorax is
   increased with age ([118]Figure 4). In contrast, comparing the
   transcriptional profile of the Drosophila heads with Gclc
   overexpression demonstrated that the differences in gene expression are
   more pronounced in young flies than in old flies (Radyuk et al.,
   [119]2012). There is data that the aging process in muscles is more
   dramatic than in the nervous system (Herndon et al., [120]2002). These
   results demonstrate a different contribution of Gclc overexpression in
   tissue-specific aging. According to our results, the DE genes in the
   thoraxes associated with aging are less than the age-related DE genes
   of the heads. On the other hand, the Zhan et al. in their work showed
   that the muscle had the largest number of age-related genes than the
   heads and other tissues of Drosophila (Zhan et al., [121]2007). The
   authors suggest that a characteristic of aging in muscle is the
   modulation of expression of genes involved in proteasomal and
   mitochondrial functions. Our data showed that the most of age-dependent
   up-regulated genes in thoraxes were associated with Protein processing
   in endoplasmic reticulum, Proteasome, RNA polymerase pathways, while
   down-regulated genes are related to metabolism and biosynthesis, such
   as Oxidative phosphorylation, Citrate cycle (TCA cycle), Biosynthesis
   of amino acids, Glycolysis/Gluconeogenesis, Glyoxylate and
   dicarboxylate metabolism, 2-Oxocarboxylic acid metabolism, Pyruvate
   metabolism, and carbohydrate metabolism. The age-related disorder of
   oxidative phosphorylation is well-known for different Drosophila body
   parts and other model organisms (Kim et al., [122]2005; Lesnefsky and
   Hoppel, [123]2006; Cannon et al., [124]2017). Cannon et al. also showed
   that carbohydrate metabolism was down-regulated in aged fly hearts,
   consisting of cardiac muscle (Cannon et al., [125]2017). This is
   consistent with our data. In addition, the age-dependent decrease in
   the expression level of citric acid cycle genes was observed in the
   heads of D. melanogaster and C. elegans (Mccarroll et al., [126]2004).

Figure 4.

   [127]Figure 4
   [128]Open in a new tab

   Heatmap illustrating differential gene expression profiles between
   transgenic (Tg) and non-transgenic (NTg) organisms: young (A) and old
   (B) flies. Heatmap colors indicate gene expression level in each
   sample, normalized to the average gene expression level across all the
   samples. Orange-above the average, blue-below the average value.
   LogFC-binary logarithm of gene expression level fold change between Tg
   and NTg organisms; p-value—edgeR's quasi-likelihood F-test p-value;
   LogCPM—binary logarithm of read counts per million; Spearman_r—Spearman
   rank correlation coefficient between Tg/NTg status and gene expression
   is presented.

   It is known that aging is accompanied by the decline of muscle mass and
   strength (Demontis et al., [129]2013). This phenomenon is called
   sarcopenia and is associated with impaired physical performance,
   cognitive function, and mortality (Gonzalez-Freire et al., [130]2018;
   Landi and Calvani, [131]2018; Van Ancum et al., [132]2018). One of the
   major factors contributing to the development of sarcopenia may be the
   decline in antioxidant protection by glutathione (Mosoni et al.,
   [133]2004; Sullivan-Gunn and Lewandowski, [134]2013). Previously, we
   demonstrated that neurospecific overexpression of Gclc delay
   aging-associated decline of the spontaneous locomotor activity of D.
   melanogaster males and females, compared to controls (Moskalev et al.,
   [135]2016). The slowing down of aging-associated decline of muscular
   function may be linked with 16–29% increase in glutathione content that
   was detected by Orr et al. in the Drosophila headless bodies (thoraces
   plus abdomens) due to neurospecific Gclc overexpression (Orr et al.,
   [136]2005).

   While the imago thorax is primarily composed of muscle tissues, several
   DE genes are muscle-specific. We observed a 1.3–2-fold decrease in the
   level of muscle-specific protein genes in D. melanogaster thorax with
   Gclc overexpression ([137]Figure 5). Using microarray analysis, Montana
   et al. demonstrated that the expression of muscle-specific proteins
   (Act88F, Act57B, Mlp60A, Msp300) and the homolog of the neuronal
   activity-regulated protein, ARC is up-regulated in Drosophila
   hypercontraction muscle mutants (Montana and Littleton, [138]2006). The
   authors concluded that the hypercontraction of muscles resulted in
   activation of actin cytoskeleton re-modulation mechanisms (Montana and
   Littleton, [139]2006). Among the differentially expressed genes in
   categories related to the muscle system process, we also observed that
   the Sclp gene was slightly decreased. Kuelzer et al. showed that Sclp
   plays an important role in programmed cell death of muscle (Kuelzer et
   al., [140]1999). In comparisons of old transgenic flies with the
   appropriate control group, we found a 1.8-fold increase in the
   expression of the genes (hbs and tum) involved in the fusion between
   myoblasts for the formation of syncytial myofibrils (Guerin and Kramer,
   [141]2009; Shelton et al., [142]2009). Sohn et al. showed that nephrin,
   the vertebrate homolog of the hbs, affects the development of the
   muscles and absence of nephrin results incompletely fused myotubes
   (Sohn et al., [143]2009). The tum gene also is conserved in C. elegans,
   Drosophila, humans, and involved in somatic muscle patterning (Guerin
   and Kramer, [144]2009). In the same comparison, the fln gene encoding
   muscle-specific protein, which is necessary for adult flies, was
   up-regulated more than two times (Henkin et al., [145]2004; Schnorrer
   et al., [146]2010). In old transgenic flies, we found an increase in
   the expression of the rotated abdomen (rt) gene encoding the POMT1
   protein. POMT1 has a highly conserved homolog in nematodes and humans
   (Lyalin et al., [147]2006). It has been shown that the mutation of
   POMT1 results in muscular dystrophy and associated neural defects in
   the mammals (Akasaka-Manya et al., [148]2004; Haines et al.,
   [149]2007). These results shed light on the mechanisms underlying delay
   of age-related locomotor activity decline in flies with Gclc
   overexpression (Moskalev et al., [150]2016).

Figure 5.

   [151]Figure 5
   [152]Open in a new tab

   The differentially expressed genes in the categories related to the
   muscle system process. LogCPM (Counts per Million) is ranged from −1.5
   to +1.5. Shades of different colors show the range of expression values
   (blue-down-regulated; orange-up-regulated). UAS-Gclc flies designated
   as Tg, Appl-GAL4 > UAS-Gclc flies—as NTg.

   The analysis of KEGG pathways demonstrated that the One carbon pool by
   folate KEGG pathway was up-regulated during aging of D. melanogaster
   with Gclc overexpression compared with age-related changes in the
   control group ([153]Supplementary Figure 2). These results correspond
   with the current paradigms. One-carbon metabolism is a crucial
   metabolic pathway that is connected with multiple physiological
   processes like de novo nucleotide synthesis and DNA methylation, and
   also provides a variety of metabolic intermediates which react either
   with pro-oxidants or promote antioxidant defense (Cavallaro et al.,
   [154]2010; Vijaya Lakshmi et al., [155]2013).

   Majumdar et al. demonstrated that folic acid and vitamin B12 can reduce
   the level of oxidative stress induced by the administration of arsenic
   trioxide. The authors consider that the prevention of mitochondrial
   dysfunction occurs due to the activation of antioxidant defense enzymes
   such as superoxide dismutase and catalase, and level of antioxidant
   glutathione (Majumdar et al., [156]2009). Vitamin B12 and folic acid
   play an important role in metabolism of homocysteine (Sahin et al.,
   [157]2003). Elderly people with lower levels of folate and vitamin B12
   have high homocysteine concentration, which is seen as a predictor of
   potential health problems such as Alzheimer's disease, cardiovascular
   disease, and loss of cognitive function (Koehler et al., [158]1997;
   Selhub, [159]2002). The up-regulation of this pathway was also observed
   by us in Drosophila flies exposed to different stress conditions like
   starvation, cold shock, or irradiation (Moskalev et al., [160]2015).

   In our prior work, we demonstrated that the neuronal overexpression of
   Gclc gene resulted in a more pronounced increase of the median and
   maximum lifespan in Drosophila females than in males (Moskalev et al.,
   [161]2016). Female longevity is observed in different taxa, but not
   well understood. In the present study, we found that overexpression of
   Gclc gene causes slight upregulation of Sod2 gene (in females), which
   participates Longevity regulating KEGG pathway. SOD acting in
   conjunction with glutathione provide a protective mechanism against
   destructive superoxide radicals and were shown to be involved in
   lifespan regulation of various model organisms including Drosophila
   species (Sohal et al., [162]2002; Tower, [163]2015). The Sod2 gene
   activity was associated with increased lifespan in Saccharomyces
   cerevisiae (Fabrizio et al., [164]2003; Unlu and Koc, [165]2007;
   Laschober et al., [166]2010), Drosophila melanogaster (Kirby et al.,
   [167]2002; Curtis et al., [168]2007), and Mus musculus (Treiber et al.,
   [169]2011). Male groups overexpressing Gclc exhibited no significant
   increases in Sod2 expression in comparison with their respective
   controls. We also showed that among the 25 DE genes found only in the
   Gclc-overexpressing female's comparison group ([170]Supplementary Table
   3), most of the genes are participants of the carbohydrate metabolism.
   Disruption of the carbohydrate metabolism and high-sugar intake are
   implicated in altering glucose homeostasis, influencing in a negative
   way the development, fertility, and lifespan in yeasts, worms, fruit
   flies, mammals (Wagner et al., [171]2015; Ravichandran et al.,
   [172]2017; Alcantar-Fernandez et al., [173]2018). It is noteworthy that
   there are 3 Maltase genes (Mal-A3 Mal-A4, Mal-A7), the differential
   expression of which were found only in transgenic females. A recent
   study conducted by Inomata and co-workers identified that the maltase
   enzymes are responsible for dietary carbohydrate changes and could
   increase the capacity of the response associated with environmental
   changes (Inomata et al., [174]2019).

   The sleep process is a physiological behavior observed in almost all
   organisms. The sleep process is controlled by two processes: sleep
   homeostasis and circadian rhythm (Borbely and Achermann, [175]1999). It
   is known that these two processes can work independently, but if they
   directly influence each other is still questionable (Deboer,
   [176]2018). In our prior studies, analysis demonstrated that the
   neuronal Gclc gene overexpression declined age-dependent alterations in
   circadian rhythmicity (Moskalev et al., [177]2016). These effects were
   accompanied by increased mRNA levels of a number of genes, which
   products are involved in regulation of circadian rhythms. It is known
   that the expression of these genes is regulated by the mechanism of a
   negative feedback loop. In brief, this loop can be described as
   follows: The CLOCK/BMAL1 heterodimers bind to the Per and Cry promoters
   to initiate their transcription. The protein products of Per and Cry
   genes form a complex in the cytoplasm with other proteins, such as
   CK1ε, and translocate back to the nucleus, where they inhibit the
   CLOCK/BMAL1 complex and repress transcription of their own and other
   genes (Yu et al., [178]2006; Lowrey and Takahashi, [179]2011). In
   contrast, in this study, the Reactome pathway analysis revealed several
   down-regulated pathways involved in circadian rhythms, such as
   Degradation of PER, Degradation of CRY, and Circadian Clock pathway. We
   found that heimdall (hll) gene encoding the long-chain fatty acid-CoA
   ligase, was significantly overexpressed (threshold of 4-fold) in
   Appl-GAL4 > UAS-Gclc flies. Recently, Thimgan et al. demonstrated that
   the expression of hll alters lipid metabolism (Thimgan et al.,
   [180]2015). Interestingly, it was also demonstrated that hll can
   modulate sleep homeostasis (Thimgan et al., [181]2015). Lipid
   metabolism and sleep are related to each other according to Taheri et
   al. ([182]2004) and Nedeltcheva et al. ([183]2009). A number of genes
   associated with lipid metabolism can modulate lifespan of model
   organisms (Bolger et al., [184]2014). Thus, Gclc overexpression had
   some effects on both processes that control sleep.

   Gclc overexpression in the D. melanogaster thorax up-regulated the rdgA
   gene. Inoue et al. revealed that the rdgA gene codes for an
   ATP-dependent Diacylglycerol kinase (DAGK) that converts DAG to
   phosphatidic acid (Inoue et al., [185]1989). According to the Reactome
   pathway analysis, the elements of phosphatidic acid biosynthetic
   process were significantly up-regulated in the thorax of flies with
   Gclc overexpression. Acting as signaling molecules, members of the DAG
   kinase family can take part in various biological processes such as
   synaptic transmission, photoreceptor transduction, and acquired and
   innate immunity (Merida et al., [186]2008). However, Lin et al.
   observed that the knockdown of rdgA gene in D. melanogaster and its
   ortholog dgk-5 in C. elegans both resulted in increased lifespan of
   model organisms. They assumed that the effects were associated with
   reduced TOR signaling (Lin et al., [187]2014).

   In D. melanogaster with Gclc overexpression, we observed 1.5–4-fold
   increase in expression of some genes involved in eggshell chorion
   formation (Cp18, Vm26Aa, Vm34Ca). Several genome-wide transcriptome
   studies showed that the expression levels of genes, which participates
   in eggshell chorion formation, decline with age (Pletcher et al.,
   [188]2002; Carlson et al., [189]2015).

   Using the GLM model “~Age + Gender + Genotype,” we compared the
   expression profiles of old Appl-GAL4 > UAS-Gclc flies with control
   groups of different ages. The old flies with Gclc overexpression
   consistently demonstrated expression profiles more similar to control
   mature flies than control flies of the same age. We found that the Dsh
   gene involved in canonical and non-canonical Wnt signaling pathways was
   up-regulated in old Appl-GAL4 > UAS-Gclc flies. Disheveled (Dsh) acts
   directly downstream of frizzled receptors and serves as an intermediary
   between frizzled and glycogen synthase kinase-3β. Dsh also participates
   in the planar cell polarity (PCP) pathway and the Wnt/Ca^2+ signaling
   pathway. The PCP pathway regulates the proper orientation of wing hairs
   and thoracic bristles in Drosophila (Kramer, [190]2016).

   Thus, neuronal Gclc overexpression in Drosophila induces
   tissue-specific transcriptomic changes in the thorax that may be
   associated with anti-aging effects such as delay of age-related
   alterations in locomotor activity and circadian rhythmicity. However,
   for more substantiated conclusions about the causative link between
   overexpression of Gclc in neurons, tissue-specific transcriptomic
   changes in muscles and anti-aging effects on the organismal level, more
   experimental evidences may be required.

Data Availability

   The sequencing data are available through the NCBI Sequence Read
   Archive (project ID [191]PRJNA474212).

Author Contributions

   AM, MS, GK, ZG, EL, and AZ wrote the manuscript text. EP, MS, LK, ZG,
   and EL carried out the experiments. GK and ZG carried out the
   bioinformatic analysis. AM and AK supervised the bioinformatic research
   and text of the manuscript. All authors read and approved the final
   manuscript.

Conflict of Interest Statement

   AZ was employed by company Insilico Medicine. The remaining authors
   declare that the research was conducted in the absence of any
   commercial or financial relationships that could be construed as a
   potential conflict of interest.

Acknowledgments